Dawn AI
Crisis detection & resources
How Dawn surfaces immediate crisis resources when it detects signs of suicidal ideation, self-harm, or severe distress. What it can and cannot do.
Quick answer
If you are in crisis right now, please reach out:
- 988 Suicide & Crisis Lifeline — call or text 988 (US).
- Crisis Text Line — text HOME to 741741.
- Emergency services — call 911.
- International — find your country's hotline at iasp.info/resources/Crisis_Centres.
You don't have to be using Dawn to use these.
What crisis detection does
Dawn monitors conversations for language patterns that may indicate:
- Suicidal ideation or self-harm intent.
- Immediate danger to self or others.
- Severe emotional distress where an external resource is more appropriate than continued AI conversation.
When the classifier fires, Dawn:
Acknowledges what you've shared with empathy
The response opens by reflecting your message back, not by dismissing it or rushing to resources. The hardest crisis responses to receive are the ones that feel formulaic — Dawn's crisis responses are designed to read like a person who heard you.
Surfaces immediate crisis resources
The 988 number, Crisis Text Line, and emergency-services numbers appear inline. International users see country- appropriate resources based on their account locale.
Stays present
Dawn doesn't end the conversation. You can keep talking. The resources are surfaced because the conversation matters, not as a way to hand you off.
Flags the conversation internally
The flag is used only for safety review and your own data record. See Privacy for what the flag does and doesn't do.
What crisis detection cannot do
Dawn is an AI. It cannot:
- Call emergency services on your behalf. Even with location data (which we don't collect), there's no automated dispatch. You or someone with you must make the call.
- Guarantee detection. Language is messy. The classifier will miss some crises and false-positive on others.
- Replace professional crisis intervention. A trained counselor on 988 can do things Dawn fundamentally cannot — including escalation paths, follow-up, and connecting you with local resources.
- Provide real-time monitoring. Dawn responds to messages. If you stop sending messages, Dawn cannot reach out.
If you're in immediate danger, please contact emergency services directly. Don't rely on Dawn for crisis intervention.
Why crisis detection is always on
The detector cannot be turned off. This is intentional:
- The cost of a missed crisis is too high to make this opt-in.
- The "intrusion" of seeing crisis resources when you weren't in crisis is small (the classifier rarely false-positives).
- The legal and ethical bar for an AI companion in this space requires it.
If you have specific concerns about the detector — false positives
that affect your trust in Dawn, or a situation where you'd want it
disabled for clinical reasons — email
[email protected]. We'll listen and document the case;
we won't change the default.
Common pitfalls
Treating Dawn as your crisis hotline
Dawn is supportive, but it is not a crisis service. The 988 Lifeline is staffed by trained counselors who can do real follow-up and escalation. Use Dawn between 988 calls if that helps; don't use it instead.
Using language that 'sounds OK' to avoid the detector
Some users learn to speak around the classifier so it doesn't trigger. Please don't. The detector is for your safety, and obfuscating means you're hiding something from a system designed to help. Use real words; let the resources surface; ignore them if they don't apply.
Safety plan
If you have a recurring crisis pattern, build a safety plan — a personalized step-by-step you can follow when crisis hits. Dawn can help you write one and reference it during conversations.
Questions or concerns
[email protected] for anything related to crisis
detection, safety planning, or how Dawn handles high-risk content.
Next steps
- Crisis resources — full directory of hotlines and emergency contacts.
- Safety plan — build your personal step-by-step.
- Dawn — privacy & data — what the crisis-flag does and doesn't do for your account.
Still need help?
Pick whichever way of getting help works best for you.
Was this article helpful?